Correlation analysis and evolutionary control of multi-staged inverted pendulum

This paper first gets the sampling values from the multi-staged inverted pendulum system controlled through the evolutionary control algorithm introduced in reference. Then the strong coupling among the state variables is demonstrated using grey correlation degree analysis in grey system theory and correlation analysis in multivariable statistics. Then principal component analysis and evolutionary computation are utilized to trim redundant information and realize satisfactory control. This scheme reduces the strong coupling between different factors and has the feature of good interpretability and small computation cost. The paper finally points out the errors in one class of control algorithm.